capture

pterodocs capture --capture model.json

Writes the loaded site model — every document, sidebar, version and permalink Docusaurus resolved — to a JSON file.

Why

Loading a Docusaurus site is the slowest part of a run, and on a large site it dominates everything else. A capture is that work, saved:

pterodocs capture --capture model.json      # slow, once
pterodocs render --model model.json         # fast, repeatedly
pterodocs sync   --model model.json --env-file .env

With --model, Docusaurus is never loaded at all.

Three uses:

• Iterating on rendering. Seconds instead of minutes per attempt.
• CI. Capture in the job that already built the site; sync in another without a Docusaurus install.
• Reporting a bug. A capture is a reproduction — it carries exactly what pterodocs saw.

Staleness

A capture is a snapshot. Edit a document and the capture no longer describes it; rendering reads document bodies from disk at their recorded paths, so a moved or deleted file is reported as unreadable. Recapture when the site changes structurally.

sync --capture <file> writes one as a side effect of a normal run, so you can keep the model of whatever you just published.